Latest posts by Andy MacLean (see all)
- Conversion rate optimization tips from some of the leading experts - June 14, 2016
- 5 powerful tips for those new to remarketing - November 19, 2015
- Bing Organic Traffic Drop – Showing As Bing.com Referral in Analytics - July 28, 2015
It’s always been a kind of obvious concept – make your web site better at selling to increase leads and revenue but it has evolved into a much more scientific and influential task than before with many really useful and dedicated tools available to the marketer to help with the process of optimising for conversions. For example in 2016 there is no shortage of AB testing SaaS providers, some of the best being SiteSpect, Optimizely, Unbounce & VWO and the price of such services is far more amenable than it used to be.
Many of the best known CRO experts and ideas are discussed in the following excellent webinar hosted by HubSpot. A transcript is also available below. Unfortunately I couldn’t grab the name of the host so he’s referred to as ‘male speaker’!
20 Questions with Today’s Top CRO Experts
[Transcript begins after introductions from 5:14]
Male Voice: Hey everyone. So I think we’re live. Welcome today to our Google Live hangout all around conversion rate optimization. So we have something very special today for everyone. We have a truly global list of world experts on conversion rate optimization based all around the world, literally. A couple of us are keeping notes, so we’re going to record this hangout so you will have it available to you. There’s a hashtag you guys can follow it’s #CROhangouts and so you can follow along ask questions and interact with people and the agenda we’re going to go through today which I think is really cool was built by everyone who is actually on this hangout. So people uploaded the questions and we’re going to answer the top 20 questions over the next 60 minutes so I’m going to be really strict with the I’m already an expert I’m sure they would want to talk for an eternity on each of these questions but I’m going to try to cut them off so we get through all the questions that you guys want us to cover and thought was important. So what I’m going to do is go around from my left to right from the list of experts. I’m going to give you a quick introduction and then we’re going to ask them to tell you where they are today so you can get a feel of how a truly globalist hangout is. So the first person is Pam Vaughan. Pam is from Health Watch, she works in our COO team. Pam has worked on an incredible amount of conversion tactics and built a conversion playbook for our blog and now she’s doing it for the website some of the things she’s done for her blog have actually increased our leads by 240 percent. Most of the times when I go and present at conferences I just steal all of Pam’s experiments because they make me look good. So Pam, do you want to give a quick introduction of where you’re based today?
Pam Vaughan: Yeah, I’m in [Unintelligible] Massachusetts in the United States of America
Male Voice: Next we have Larry Kim. Larry Kim is a CTO of Wordstream also one of the people I see published in lots of lots of content and an incredible amount of content. A list that’s always, full of lots of earnings. So he has the best known person for paid acquisition, an expert in CRO, a phenomenal marketer, and also someone who has gotten up really early in the morning to be good enough to join us today. So, Larry do you want to give us a quick introduction and let people know where you’re from.
Larry Kim: Sure, I live in Cambridge, right near the HubSpot office but I’m actually dialing in from Sydney Australia this morning and that’s why it’s so dark here
Male Voice: Awesome.
Larry Kim: it’s like 3 AM
Male Voice: 3 AM, so that is commitment to this hangout. Thank you for doing that Larry. Oli Gardner, co-founder of Unbounce, I’m sure everyone is familiar with Oli probably seen more landing pages in his time than anyone else. Renowned speaker, speaking, spoke all over the globe, being kind enough to speak in at in Dublin, where I’m actually based and has a blog on Unbounce, amazing things on right version tactics. What Oli does is really cool. They do a show where he tears down landing pages as well. Oli do you want to give a quick introduction and let the people know where you’re from.
Oli Gardner: Yes I’m in beautiful Vancouver British Columbia Canada.
Male Voice: Awesome.
Oli Gardner: and it’s not that early, well its 9 am. That’s early for me.
Male Voice: 9 am, not too bad, not competing with Larry. Peep Laja
Peep Laja: Laja
Male Voice: [Unintelligible] I think I just got it wrong. Peep runs conversion. Co-founder of Conversion XL, entrepreneur, over ten years of industry experience. The best thing I could say about Peep that happens anytime anyone wants to ask me a question on conversion or CRO I just say go read his blog. I’ve stolen multiple things in his blog; it is definitely one of the best resources on the planet for all things conversion. Peep do you want to give us a quick introduction to let us know where you’re based today?
Peep Laja: yeah, I’m in Austin Texas. Also, the United States of America.
Male Voice: Another United States of America. And then last but no means least is Rand Fishkin. I don’t think you need too much of an introduction. Founder, former COO of Moz. Author of tons and tons of books, co-author of lots of books. Co-founder of Inbound.org and an industry speaker across the globe. Rand, where are, where are you based today because you usually travel quite a lot.
Rand Fishkin: I do, I’m actually in Seattle Washington. My hometown, which is kind of shocking. It is in the United States but my hope is and maybe Oli you can work on this, if Trump happens to win the election maybe Canada can invade.
Oli Gardner: Yeah we can absorb, we can absorb you, and we’ll just do that.
Rand Fishkin: That would be, that would be really appreciated.
Male Voice: Cool, so there are our speakers. You have to pay a lot of money to go see these people speak. So we’re giving that to you for free. They’ve all been kind enough to give us some time. So we really appreciate that and want to say thanks for their time. Let’s get right into the questions. So very first question I think is a really great question to start off with so this for, for you Peep. We’re going to go to you. How do I create a CRO process? How do I get started with that?
Peep Laja: Well first thing to understand is what do we need the process for? And the goal of what we’re doing here is either make or change, or test effective changes in our website. So a good structured process, CRO process should help you answer that question. The question of what should we test next. What should we change next on our website so a structured process will tell you what the problems are, where they are on your website? Why are these problems even problems and it will help you, once you identify the problems, turn them in to a test hypotheses which means solutions to these problems with an A/B test should be a solution to a problem. So a CRO process should focus primarily on identifying what the problems are so you want to look at the analytics to figure out where the traffic is dropping out. You want to analyze a qualitative feedback to understand what’s keeping people from signing up. Why are they leaving your website? Why aren’t they buying more stuff? You want to use all the tools available to you to figure out what the problems are, so your structured process should be around identifying problems, turning them into test hypotheses and then running tests and I think within this three minute time limit can’t give you a better overview.
Oli Gardner: Now, I think people should check out your Research Excel Framework.
Peep Laja: Yes they should. Thank you, Oli.
Male Voice: So, wait, give that to us again. What was that? The[…]
Peep Laja: Research Excel
Male Voice: can you give us a little bit of insight of what that is? It is a way to[…]
Peep Laja: It’s a framework that I’ve developed that teaches you how to do it. So if you go to conversionexcel.com in the top menu there are free guides. A free CRO course is a full course on how to build the process.
Male Voice: Awesome. Anyone else wants to jump in there. We have a little bit of time on how you would create process prioritized. Rand, you have something?
Rand Fishkin: The one thing that’s worked for us pretty well is to identify the area that we’re trying to improve and then talk to two groups of customers in particular. One are people who are right customers but did not end up signing up so meaning they match all the criteria of people we think should have loved the product and wanted to sign up for it and buy it or just sign up for the email, whatever. Do the conversion event and then also talked to the people who did sign up for it and ask them the same set of questions. Things like why were you interested? What initially brought you there? What made you decide to investigate it? What made you decide to convert or not convert? And then you can look at those objections and turn that into the process of things that you test on the landing page. Answering those objections, overcoming them that seems to be a really powerful process that you know that I’ve stolen from experts like Oli and Pam and Peep over the years and it has worked well for us.
Male Voice: Really quickly, how do you ask those questions? Is it like through an email survey or websites surgery?
Rand Fishkin: We’ve done email surveys. We’ve done in person. So, a lot of conferences and events, I talked to folks about this or we’ll bring them into the Moz-Plex and ask them in person and phone calls work well too.
Peep Laja: The email is the way that is the scalable approach which you know survey hundreds and hundreds of people fast. So, interviews are not very scalable, but you know of course you get richer information if you do in person interviews.
Rand Fishkin: Do you find Peep or other folks, do you find that when you, I find that when we get like five or ten answers those tend to be very similar. The same objections come up if we ask ten people or a hundred people or a thousand people.
Peep Laja: So…
Rand Fishkin: Get more and more.
Peep Laja: You can identify the top issues with a little smaller sample size but if you interview only ten people what you don’t get is you don’t understand the severity or let’s say that frequency of the issue. So if people worried about the price, security, and I mean whatever a third reason. So, like what is the scale of different. So if you survey five hundred people you understand that, “Oh my God, pricing is at times bigger issue then fear about something you know.
Rand Fishkin: Right, got it, got it.
Male Voice: so we’re like I guess the next question is really great fall on [Unintelligible]. We’re going to go to Pam for this. What are some of the top CRO metrics a person should really care about when they’re trying to create that first process? What are like the next stage [Unintelligible] towards?
Pam Vaughan: So, I honestly think it depends on, you know, on the goal. On the problem you’re trying to solve. What you know, what you’re trying, the problem you’re trying to solve will have different metrics depending on what it is. So, I mean the go to metrics. You know when they talk about ‘Conversion’ Rate Optimization with no conversion. Conversion rate which you know also depends on traffic. But you know if your goal for example is to improve the number of marketing qualified leads that you’re generating from your blog. You can’t just look at raw conversions, right? You have to look at the qualified conversions that come from the blog. So things will change depending on, you know, what products and experiments you’re running on, you know, if you’re running an email experiment you’re going to be looking at, you know, the true nature of your emails and, you know, referral traffic from your emails and things like that. So, really I think the main thing is to focus on the goal you’re trying to solve and hypotheses that you formed for that experiment and then what make sense to actually measure to evaluate whether you’re your hypothesis is true or false. And another thing I’d like to point out was the big role that traffic plays into all of this. I think traffic is a huge huge metric and particularly traffic sources.
Male Voice: Yea
Pam Vaughan: No conversion optimization, you know it’s really important to pay attention to the traffic. If you’re trying to optimize a specific age to understand how people are getting to that page. People behave differently depending on the source of traffic. I mean, you’ll notice changes in conversion rates based on that. So, you know things that we notice is that you know traffic typically converts at a lower rate than something like organic traffic because visitors from organic traffic are much more qualified typically, while they were searching for something that you rank for and they ended up on your page. They might be a little bit more qualified then people that you might acquire through paid traffic. So, I think it’s really important to understand the sources of traffic that you’re dealing with in any given project and understand how best to do, you know, cater to those sources of traffic.
Male Voice: [Unintelligible] Let’s move a little bit further into the process of someone’s creative process. So they understand the metrics and the first thing they want to do is like redesign one of their landing pages, redesign one of their site pages to try to improve conversion. How do you think they should go by approaching the design of that page to improve conversion? Oli, I want to throw it over to you, a man who seen more online pages, site pages than anyone else.
Oli Gardner: Well the first thing you shouldn’t do is like jump the gun and jump straight into designing your page that you have to have a process for this too. The first thing should be copy. Alright remember copying forms design, not the other way around. Which speaks to a common question and you know who’s heard “The ugly online pages convert better then beautiful ones,” right? This is it. People say all the time but it’s not true. The reason why ugly pages do convert is because that person didn’t have design skills and didn’t worry about it and they all focus all of their time on copywriting. Which is why it’s persuasive, which is why it’s impactful? Could be better with extra design wrapped around it? Most likely, because that would just be an even better page. Conversely you know if you’re a high price brand, Prada or whatever you have to look exceptional because that’s part of your visual identity. So they can’t get away with it, but a lot of businesses can. So focusing on copywriting is the start of the process. Once you have that and then your information hierarchy you tell your story in the right order then you can wrap a design around it. So we’re going to design now, you have to realize that traditionally no one has been taught how to design for conversion, nobody, right, graphic and web designers have been taught typography, grid systems, color theory, and some aspects of it. But really it’s been designing for you x or designing for winning awards or whatever it is but not for conversion which is why I recently wrote a e-book called “Attention to Design.” It teaches you how to use design to focus attention on what you want people to be doing and give you a business advantage through design and interesting story [Unintelligible] you asked what it was like speaking in Dublin and when I got there I was checking in the hotel. I got my key card for room 324. I went to the elevator. There was a lady there ahead of me. We’re chatting, I let her in the elevator first, she puts her key card in the third floor, and I’m also on the third floor, so I don’t have to do anything. Now when she exits, I exit behind her. I’m following her down the corridor which is really creepy and awkward and it’s not until I get two doors down that I see the numbers are going in the wrong direction. I’m not going to my hotel room. Here’s the actual, this is the sign that made me go the wrong way and it’s not until she enters her room that I go (gasps). I turn and run away because that was the second door. I’m looking like a pervert and so I ran an experiment to see how many people did this and I had 33 percent went the wrong way. So I changed it using design principles, contrast, grouping the numbers next to the arrows a bit of separation and I through three experiment being 5 second tests, I managed to get a hundred percent of people going the right way. Which so if you can identify a design problem and you understand the principles, you can solve the problem through design. It’s a fascinating way to be able to, you know, look at things and actually if you want the e-book, here’s the link to it, bitly.com/add-ebook.
Rand Fishkin: So now we can add to your title Oli Gardner, hotel sign designer.
Oli Gardner: Could be.
Rand Fishkin: Your prestige is rising as we speak.
Oli Gardner: I’m tired of solving the world’s worst problems.
Oli Gardner: I love seeing these real world problems; through showing with a five second test this is a problem and then fixing it. But going back to the copy part, clarity is the most important of conversion. Which is why copy only can convert. This is an equation which identifies how good the clarity is on your page. It’s broken down to seven sub equations that then let you know where you have a problem with its immediacy. Can people get it right away? Is it readability? There are seven of them and it’s, hit me up on Twitter, I’ll share that with you. It’s a great way of figuring out if you have a client who is like
Rand Fishkin: Is there a URL for that?
Oli Gardner: Not for that. It’s brand new. I’ve only talked about it once, so I don’t like talking about it and
Rand Fishkin: So, we have to go get the recording, take a screenshot and then post it.
Oli Gardner: Well tweet at me and I’ll share something.
Rand Fishkin: I’m on it.
Male Voice: What I’m on, is one thing you said there Oli that I would be interested in. Well everyone thinks that it seems that you start every, maybe it’s not every test, that you always start with copy and linear design, you think that’s always the case that you would always look a copy first and design and I guess it follows that is what people this is the best path copy vs copy test and then design vs designing. You change one of those things at a time or not?
Oli Gardner: Well ideally you start with copy, right? That’s because the best process because then you understand what treat you need to present that copy. Otherwise if you start the template you’re just filling in holes. Square peg, round holes. Sometimes you know it has three bullets, I have to write three bullets. That’s not appropriate, maybe you need twelve, and maybe you just need one. And when it comes to testing that should be informed by your research. You shouldn’t be well, let’s think about a copy test today or let’s think a about design test. You need to identify there’s a design problem, use design to fix it. If you identify there’s readability or a clarity problem use copy to address that.
Male Voice: Ok, we are going to the next question. We’ve got here one of the questions people really want to get answered was, what are the top three CRO myths that are out right now? So, what are people really getting wrong around CRO that they believe it’s kind of fact. We are going to go all the way to Australia. So, Larry you can pick up this one.
Larry Kim: Sure thanks. How much time do we have here, I’m just kidding. I think if it was just the top three of big myths on CRO or big mistakes, places where people get tripped up. One of the would definitely have to be this idea that smaller changes are you know generally to smaller changes and they often actually don’t exist forever. A lot of times, you know the reason why something works so great is because it’s new and then later it doesn’t work so well because it’s no longer new. Another thing that people will often get wrong is this idea of the diminishing returns and kind of a theoretical maximum conversion rate. You don’t believe it’s possible to go from one to three percent conversion rates. You know a little bit harder to go from three to five, much harder to go from five to eight. You see what I’m saying? You can’t go to a hundred percent and beyond usually on and so and I just know when thinking about your strategies, sometimes people maybe overestimate how high we can get this thing, you know, because the kind of low hanging fruit. Another kind of area I think people can get tripped up at is decimals an increase in quantity doesn’t necessarily mean an increase in quality so just be mindful that you’re not just making changes to your marketing stuff to go more unqualified people. You always have to be checking against the marketing, qualified leads including Leadin because there’s costs associated with following up on leads. You don’t want to inundate people with crap leads. Anything to add this, anyone?
Peep Laja: I would like to add a few things here. I think the biggest myths in CRO is that people think that CRO is a list of tactics, a list of, there’s one great CRO hack around the corner that will change everything. I mean, if you think of CRO as just a list of tactics, so I’ll just find that blog post put 100 tactics, you know my conversion will grow whatever. That’s BS, you always need to start with a process to identify what your actual problems are because if you copy your competitors, you’re copying their solutions to their problems. Well actually they probably copied somebody else and copying competitors, copying market leaders also BS don’t do that and the myth of CRO is figuring out what works. So we’ll just brainstorm great ideas and then let’s test them. That is the stupidest way to go about it. I mean an A/B test or any change you make on your website is about addressing a known identified problem. So, focus on identifying what those problems are. Don’t rely on the next great CRO hack.
Male Voice: Well does anyone else want to jump on the answer before we move on? No? Ok, we got that one covered. So this is what I think is one of the most important parts of a good CRO process and there’s lots of frameworks. I want to go over to you Rand to get your thoughts on this. How do you score and prioritize experiments within laws or how is this done really?
Rand Fishkin: Yeah so when we run experiments, we are generally seeking the same thing. We’re a SAS product with a software subscription. So we’re almost always trying to increase the number of people signing up for a free trial or signing up for a free Moz account, a community account. So for us scoring those is very simple right? We can essentially say that given the number of visitors that landed on the page, or that came into the funnel at all, is broken down by source as Pam noted, that we can then look at the percent of folks who signed up. The only thing that’s critical for us in that metrics and that scoring in addition to just the raw numbers is the performance of those folks once they get into the product. So in a SAS product right, it’s not E-Commerce where you’re selling something once or hoping to sell maybe you have some real recurring purchases in the future. You’re looking for that recurring revenue and so for us one customer stays with us for two or three years is worth far more than ten customers that stay with us for a month or two months. Each turn rate is kind of a killer you don’t want to get people into your funnel that are not right for your product. So when you look at the performance of the conversion rate but we also use a methodology to look at the first three or four months of performance of the customers who signed up just to qualify whether they have a turning curve that’s you know similar to better than worse than our existing customer sets and I think this is important for anyone who is considering a signup process. For example, one of my pet peeves, one of the things I hate most in the web world right now is the pop ups that asked for your email address and are like do you want to, you know, learn more about SEO and then there’s a big button that says “Yes” and a smaller button that says “No, I’m dumb and I want to stay dumb forever.
Oli Gardner: And then come back up
Rand Fishkin: Well I hate that. It just drives me bananas right and I think that one of the things that I would say is very important to test is not just did you get more email signups because I’m sure you do get more email signups from that account button but how many of those people are long-term subscribers? How many of them are engaged subscribers? What percent of them hit unsubscribe? What percent potentially hit report spam and affect your email deliverability and email reputation sender score I think those things are critical to consider so when your qualifying a test, remember to think beyond the point of conversion to recidivism, retention and lifetime value.
Oli Gardner: What was that word you used?
Rand Fishkin: Retention and recidivism; meaning you come back and do the same thing again. Right, so like in e-commerce someone buys one pair of shoes from Zappos. Do they ever buy again from Zappos?
Oli Gardner: So I learned a new word today and [Unintelligible] almost learned a word which is Peeps last name.
Rand Fishkin: He almost learned Peeps last name.
Male Voice: Its only 4 letters
Pam Vaughan: Yeah I’d love to chime in on the priorities things. Yeah on our team we use the PIE framework. PIE stands for Potential Importance Ease. Essentially you know we’ll have you know people come to us all the time with ideas for optimization projects we come up with ideas ourselves based on your digging through our analytics and determining what our problems are based on what are our company goals, and then we take these projects in and we give them this PIE score. Essentially know how much potential improvement could the result of this project or this experiment be? The importance, like how important is this to our business? How important will the improvement that we make be to our business? And then E means how complicated is it to actually execute this project. Is their kind of stakeholder, political barriers you might come in contact with? Are their technical implementation, road blocks that we might have to overcome? And then each of these categories gets a score from one to ten. You know ten meaning does it have the most potential? Is it really important and is it really easy to execute and then we average these scores out to get an overall high score. You know for all the projects that we’re evaluating and we prioritize them based of those scores.
Rand Fishkin: Pam do you have a resource or a link to where we could see that you listed out?
Pam Vaughan: Trying to think of [Unintelligible], do you remember where this PIE framework is from?
Male Voice: The PIE framework actually comes from a conversion, a really great conversion company based in states, that I’ve gone I know them really well.
Peep Laja: It’s in Canada.
Pam Vaughan: Yeah
Male Voice: They have a great resource section and great stuff in their blog about this.
Peep Laja: Yeah lots of preservation frameworks. There’s an ice model, similar stuff from the eighties and then their binary models like if you want to create a culture of optimization within your company and you want to score let’s say an A/B testing ideas you want to add criteria’s in like does this idea come from user feedback? Yes or no. Does this idea come from I don’t like does it have support from analytics? Did you look at this thing? Is it above the fold the change did you add or remove anything? So all these binary criteria’s so next time when people come to the table with an A/B testing idea that the just brainstorming in the bathroom is the no user feedback no analytics data blah blah blah. Ok, next time actually I’ll go out and get some data so with the privatization framework you can influence the way people think.
Male Voice: Yeah I think the thing with the PIE framework is you can get a PIE score if you’ve done a lot of those things. You don’t you wouldn’t know how to find it.
Peep Laja: Right you are estimating potentially I think this is a seven and so on.
Male Voice: So let’s get some healthy debate going if we haven’t already. Rand mentioned something that is hard topic in the market and people have passionate feelings about. Which are pop ups. So, Oli, let’s get started with you and then people can come at you. What are your thoughts of a local expert?
Oli Gardner: They suck and they’re great, right?
Oli Gardner: In truth all they are is an evolution of the mechanics of how we request an action that people online. Does it make it good no or not always like if they offer something you really want their amazing. If it’s the right place at the right time we have some amazing success with specifically, like typically they’re converting one four percent but we had digital agency day it was an event kind of like this show up and we have recordings for it so it is an event not everybody can attend so on the exit there is like, “Hey put your email in and we’ll give you all the recordings.” That converts thirty two percent which is unheard of because it was super relevant and helpful and the truth is they’re part of our future. That future being conversion automation. Well what does that mean? It’s a means taking all of the conversion opportunities that exist on every page on your website landing pages and having an evolved to create experiences that are delivered in a responsible manner. Responsible being the important word so it’s capitalizing on those conversion opportunities but perhaps more important it’s discovering the parts of your pages of your marketing that need to be protected. Where do you need to put this walled garden? If you can find the data on these things then be like well this page can be really successful in requesting that but this one doesn’t ever put anything there because that breaks part of the conversion chain and you have to be really careful that certain pages, any interruption ruins that experience people leaving don’t come back. So it’s not just about capitalizing everywhere it’s about figuring out where you have to put some kind of barriers to stop that from happening. You develop a conversion opportunity map and start from there because every website and landing page has conversion opportunities and every conversion opportunity needs a call to action sometimes that’s whether you need to do nothing or not and it’s a responsible use of technology, psychology, copyright interaction, and design. Now, it’s interesting. I was on my phone where they play a game Genies & Gems, [Unintelligible], and I got in and here’s my progress, right. I’m like on level 50 and then a pop-up comes up and says connect to Facebook and save your progress. That’s a blatant lie because it doesn’t matter if I do or not. Your progress is always saved and I didn’t do 50 levels in one day. This is overall course of a week. It’s lying to me to try and get me to share this on social. That’s when he responds to the least of you know interaction, copywriting, and psychology. Above all with the other things you have to be delightful and there’s always a way to make something delightful, if you demonstrate empathy for your visitors.
Rand Fishkin: Yeah I heard there was a game. I can’t remember what the game was but it’s like a tower defense game and you got a little power up if you did the share, right? So they give you like a free something if you did the share so it wasn’t a lie and you had your choice right and even I didn’t connect it to my primary Twitter account and I created a new one and tweeted it out from that random account. I needed that power up. It was less of a dark pattern; I think you know you got to be careful about abusing trust.
Peep Laja: Most people don’t take action on the first visit. Do you like to subscribe or whatever. So if you’re working hard to spend money or maybe at the end of work and SEO whatever to do drive traffic to your website and they come to look at the stuff and they leave there’s like oh my god I’m out of all the PPC budget but if you have that pop ups you can capture some of the traffic remarket it to them, get them to come back because the difference between static options of put your email here always visible vs pop up it’s an amazing difference like seven eight nine ten times difference so it’s hard to, it’s easy to bash pop ups. It’s always in like such a you know a guy lying down on the floor let’s kick the ass; you know it’s like sorry I mean it’s so easy to bash it but it is a fact give. If you’re not and you know well as you said as always said if you’re not on the ethical about it you know
Rand Fishkin: Yeah I guess my point is broader. It’s just that if you’re going to engage in pop-ups or especially more aggressive ones right with negative messaging, just make sure that you’re testing how those emails, that the sign ups that you get perform versus the ones that you got in a less aggressive manner.
Pam Vaughan: I share your hatred of those pop ups. Bringing that “No I don’t want to be a better person” Just what I think is that I think there’s a way to do that is not annoying and is you know taking your users experience in the consideration on now putting. For example though we have an exit popup on our blog for a subscription pop up and you know it’s an exit pop up so people get it when they’re about to close down to the browser but we also had it on a timer so that we’re only showing its people who have been on the blog for a certain amount of time because you want to only encourage the people who are enjoying reading your content to do to subscribe has been another. Also how to text about it my does exit out of it we will show that again them for certain short period of time if they come back they’re not going to do that again and try and exit out of it. It’s smart that it already subscribed. We’re going to show them to subscribe you know that’s just common sense. So I just think there’s ways to just be mindful of your users experience when you’re designing these pop ups and not overdo it and you know like you said be mindful about it.
Oli Gardner: So the future of this is where machine learning is going to come in and that’s going to do two things when it gets really smart first of all let’s think. So you’re on a page it will test automatically; exit intent, scroll intent, and time and collect all of these things for you and say on this page this performs best but even smarter. Hopefully we’ll tie back into like a cohort analysis and go but these people didn’t stick around two rounds, or these became high lifetime value and you can when you can connect those things like this interaction model actually helped to get better quality customers. When it’s going to get really smart, really intelligent and it will. It will start to undo some of the damage some of the harm that’s being done because it will be more intelligent that’s the reason.
Rand Fishkin: Is there any software like that?
Oli Gardner: Not really [Unintelligible] exchange
Peep Laja: [Unintelligible] gotten really smart it’s just the down under the hood stuff is impressive there that’s the top of the line it’s you know it doesn’t do the kind of looking at the code works by a man in the back in like the stuff yet but I mean might be coming, I don’t know.
Male Voice: Guys I got to cut it off in my [Unintelligible] way to move onto another question. So I appreciate all the great information around that. We are going to head all the way back over to Australia, to Larry. I think one of the things people will be interested in was when it comes to you. So Pam mentioned earlier on about source of the traffic being very important and in particular sometimes converts less than a lot of the other sources like organic direct. What are some of the key CRO lessons you’ve learned specifically when it comes to like [Unintelligible] or social advertising?
Larry Kim: Sure well first of all of I don’t know like about that statistic because my page channels are like by far our highest converting channels for my own business so maybe we should have a talk with whoever’s running your [Unintelligible] channels [Unintelligible]. There’s a ton of lessons of CRO that you can get from PVC so I’ll just give me a couple crazy examples. So generally it though one of the most important key metrics that a CRO should be thinking about in terms of PVC is click through rate alright. So click through rate is ridiculously important because what we find it when I analyzed like billions of dollars in ad spin. What I find is that if you can raise the click through rate your ads. If you double your click through it will double your conversion rate so not only that it’ll actually, you’ll get more clicks because if you have doubled the number of people going through so like a four times increase. So what the heck is happening here? Generally, if you can get people really excited about kind of whatever it is you’re offering that excitement generally will carry through to like a purchase or elite leaves a flow like downstream. So click through rate I think it’s also really important because I think in some ways it’s even more important than conversion rates because conversion rates are biased. Conversion rates, its only telling you like in your offer, it’s telling how interesting that offer is to the people who kind of self-selected into wanting to learn more about that offer like they’ve actually click through there. You could be in a situation where you know if you have a very low click through rate but a very high conversion rate that could tell you that well we’ve got a really niche market basically. We’ve got kind of a really appealing offer that it feels like a very small amount of the total population of people who actually buys this thing. So like I think PVC I go in terms of the click-through rates can give you a ton of perspective in terms of like you know not just sampling how good your offerings amongst people are interested but also the people who you know like how appealing it is globally speaking and if it’s really not doing well from a click through perspective that should be you know maybe a clue to maybe consider diversifying your offers or changing it up a little bit. Another crazy thing about PVC is that our advertising in general is that brand awareness plays a massive role in biasing people and thus impacting conversion rates in a huge way. I’m talking like three four five times that gets, I mean you can prove this to yourself guys just go to your analytics, look at like your conversion data and look to see like what’s the difference between like people who convert that are repeat visitors vs new visitors right. People who are familiar with the brand or vs people who are not previously familiar with the brand and you’ll find that the conversion rates are ridiculously higher typically two three four five times higher for the repeat visitors and so that’s basically what advertising allows you to do is to go after those people and familiarize themselves with the brand so that when they do search for things that are related to your product series following they will be biased though want to go with what they know and this displays a huge margin rate optimization and of course it would be crazy to not mention remarketing. I think we’re marketing is a massive CRO hack it’s located on the thing but you know it’s changing that one opportunity to get the sale into like hot potentially. You know thousands of opportunities to get people to remember what it is you do and also convert at a later time you know people are just busy may have had to juggle some other task and they didn’t finish the thing they were hoping for them and I think remarketing is such a game changer and that you can kind of just have a second or third or fourth chance at a conversion.
Male Voice: Awesome. Thanks a lot for that Larry. All right let’s quickly jump into the next question because I’m being terrible timer.
Rand Fishkin: I know we are short on time but I just want to mention one quick thing on that what Larry mentioned and Pam mentioned about paid versus organic traffic sources. So, Larry I agree with you that our paid traffic sources of models also convert better as a whole then our organic ones but when it’s one and one when it’s exactly the same. So for example, our paid Twitter campaigns, you know two landing pages that trying to convert people to a product don’t work as well as our organic tweets to those same ones our Facebook ads to a product landing page don’t work as well as a Facebook organic post to the same thing and the same is true in organic vs paid SEO vs PVC. So someone searches or you know cured research SEO tool we have an ad for it we also have an organic tends to convert. I think only slightly but a little bit better than the paid traffic does so it isn’t available for sure but one to one organic.
Larry: I just think you need to be more picky. Who says you have to remarket to everyone who’s visiting your site. Why not apply some demographic filters and only market to the people who have purchased business offers in the last 60 days or something like this. Something like I certainly don’t want to be in a situation where you know you’re paying money for stuff that’s not doing you know better than your organic stuff. It should be many times better and then you know that’s just like a targeting issue I think.
Rand Fishkin: Ok yeah, I should look at it.
Male Voice: Ok we’re going to go to [Unintelligible]. I think this is pretty good question is, How can you, is there such a thing is like over testing and over optimizing and can you just go a little too far
Peep Laja: And I mean you got that your conversion rate is like hundred percent and everybody buys don’t want that. There is no thing is over offers attention I want you all over testing what you can do with whether it is a problematic issues if you don’t know how to run tests so easy if you stop tests too early or if you have lots of tests running at the same time that are interacting with each other meaning biasing the results that can happen so it’s not about over testing I mean every single day without the test running is a day wastes so you need to be testing all the time provided that you have the traffic is a lot of small businesses shouldn’t be testing at all like I mean broadly speaking if you have you know if you’re less than 500 transactions a month you shouldn’t be testing at all if you have less than 1000 or maybe you can do one test a month or so to speak so over testing if you invest early as possible but if you do it right if you know what you know if you know how to run tests there is no such thing is over testing or over optimizing I mean as Larry was saying I mean the peak of how far we can get I mean it’s the mountain is very tall you know we should be continuing to climb.
Larry: Yeah but here’s the thing out so that the you want to turn your donkeys into unicorns you don’t want to turn your unicorns into donkeys you know it’s like that it’s kind of like when you get something that’s pretty Good you know then you know you’re not every test is going to go your way and so like that the works you start off with kind of like our the higher the likelihood for success and converse is that the higher least turn off the higher the probability for failure and so that you know like I can draw a little curve where there is kind of like this optimal point where you know it’s like that this it’s pretty good let’s not turn our unicorns into donkeys.
Peep Laja: well I mean if you’re talking about opportunity cost instead of optimization with it could be doing something else then yes if it’s just you know lack of human resources whatever it can be can happen but then again that we have it is right that means we have a one-percent win it’s booking.com gets a one percent win oh that’s so small one percent Oh actually it’s another hundred million dollars you know
Male Voice: alright cool so we are at kind of got to switch gears a lttle bit towards blogs. so a lot of people who are on this are trying to figure out how to get some sort of conversion payment for the blog and get something kind of go back to your blog one of the questions the what he wanted to know was for you Pam, you know every experiment as possible on a blog is that we have a question on have any CTA should be on a blog post or page. I guess we’re trying to get to is like what is the right way to try to convert some of your blog post.
Pam Vaughan: so I don’t know if there’s a specific number. I actually, in testing and I’ve done one of the things that has really stood out to me is but there’s a definite correlation between well conversion rate and the placement of the CTA on so I’ll just give you an example so we have on our market why we have really long blog posts on some of them are three thousand words long we found essentially that you know all we had was that end of post call to action like they ever see TJ because I am first we ran some heat loss and roll that our love of your life people aren’t even making it fun they weren’t even like hitting the CTA where banking on that when we started introducing is higher having a post on action call it I call it an anchor text CTA basically within the first few Paragraphs you know download this eBook or this template to learn how to write a press release for this press release tablet we basically doubled our conversion rate from our blood by doing that I what’s interesting is that a lot of the work that I’ve done a lot has been on historically and optimizing our old post but continued to get a lot of traffic I mainly from search but don’t come hurt very well and I just started doing that I just naturally started to add these you know anchor text et is near the top of the post and what mattered was basing it off of the search terms that people using to get to the post so in that press release example we have this life was already a press release I it ranks really highly for some really high volume key words around how to write a press release to the press release template and when I want an inversion optimize that post using those keywords I added and I any of the top and it essentially doubled the conversion rating when we apply this across all of our historical blog posts I’m basically matching the keyword that they were using to find the blog post in search with you know you think that saying turn in the CTA for the offer that was very relevant to that a lot of post I’m that’s what are your conversion rate kind of him through the roof and I think it was it was a combination of the relevancy of the keepers that we were targeting I’m going to post but also the place then um because you know originally I thought it was just had to do with the keywords and then I ran a study that basically look the breakdown of where we were coming from on the post level like which days were actually generating all the way and they were like on at like the end of the post call to action only generated like six percent of the total ease of those posts they were all being generated from those calls to action here at the top I so I mean I think the takeaway here is that I there’s a ton of different CTAs that you can use on your home um you know we trying to balance out so if it’s a really long post like we’ll have that you too ok well how the end of OCT a you know not overdoing it but I think it’s the latest how long the posters as well because you want to give them opportunities as they are reading through to convert their ready for it.
Oli Gardner: Yeah we’ve had the we did exactly the same thing recently not with we always have a little text links to relevant stuff but we r you we had been at the bottom say pathetic performance we moved him up there’s a redesigned slightly and made them super relevant to that paragraph we actually have it so when you scroll down to it it will kind of light up like a panel as you move over it goes hey how about me and it converts so well but you know when you think about how many CTAs it’s different from a blog post of landing page landing page one that’s obvious and although if you need to you can have a safety net sea-tac maybe that’s on exit intent or ideally after the conversion of the confirmation page you can get the other thing maybe you’re dying to get but on content or we found because you look at your top 20 organic content pages i bet half of them have outdated content after them don’t have an actual CT and then actually asking for anything we we did this she type was a landing page we on that right we’re number one Google we have had it for seven years and it’s this printer seven year old piece of content with lots of internal links to other places and ran can speak to this much better than I can about like say for a legacy page like this is what we also what we did first so we took out we analyzed the google analytics what is this we know our best conversion path people have to hit the landing page templates page that if they hit that they’re much more likely to convert you on this page we stripped out every internal link in the content we took out the side and we just added the CTA to say hey check out our templates hundred and seventy one percent increase in people going where we wanted them to go right and by Then so my question I guess for Randy’s like what do you think about this page has existed for a long time and it has a lot of connections into the site if you remove those content links I don’t know what the impact of that might be I mean the result was amazing for us but you know I don’t know
Rand Fishkin: yeah typically I mean on a site on a site your size with the number of links and domains that are pointing to you and the quantity of content that that one change even to a powerful page on the site is not going to have a massive impact it could affect that single page or maybe a couple other pages I was like to but I would worry far less about that and far more about the conversion rate in general so I think that hundred and seventy one percent you know if you move from ranking three or four for a few of those internal pages that you were pointing to to five or six I don’t think I sweat that in your shoes right and you’re really not going to see much more movement than that given how powerful the unbound cited
Oli Gardner: yeah it was interesting because we took away the links it went up maybe twenty percent which is great the sidebar we have the CTA went to have the fifty percent then we took away the side after 170 times anyone to each one of those little things make a little difference and it was great on landing page this is very different you know you should have a single CTA likes it before focus on your copywriting I think the three most important skills for a modern marketer are writing coding and public speaking if you can do those three things really well you know it marketers need to get more technical now it’s just like any ball back in business if you have to rely on someone else then you’re going to be slowed down this I think this is why growth hacking became popular because that it gave people an identity and our technical marketers I prefer that term didn’t have an identity that would look down upon by the dead community because of your marketer what do you know about that a lot of people our technical and they can I i was a coder first so I can I can build things without anyone’s help sloppily and I wouldn’t want someone to look at my code yeah but I’ve been afraid to do those things and also what you’re asking for on your pages of important there’s a thing called I call conversion hierarchy which is the which is the conversion journey from first touch to an activated lifetime highlights value customer and there are jumped in this and if you ask for too big of a jump people won’t convert you have to do smaller jumps but there are exceptions there are accelerants you can use video is a great accelerant because if someone watching your video you can push them up your way quicker and that’s how you kind of do some of those things especially its a workshop that shows your product or something they’re great accelerants that you have to if you think about your conversion hierarchy and start planning how you can move people along it.
Male Voice: Cool. Awesome I’m going to try to get two more questions in before we have to wrap up. So we’re going to go across to Ron. Any conversion but a lot of a lot of focus on SEO this question is should be more optical about on page SEO factors? I’m not sure if I pronounced that word correctly.
Rand Fishkin: No we should not be more skeptical about on page SEO we should just move over to the right forms of on-page SEO I think one of the wonderful confluences of SEO and CRO is the fact that today a high engagement rate a low balance rate and high browse rate are strongly correlated in my opinion actually have a causal relationship with high rankings in google and I think google is looking at engagement and they are looking at how traffic performs on a website and they are looking at bounce rate and what they call pogo sticking where you bounce back to the syrup and you click on someone else’s result and for that reason we need to do an extraordinarily good job of converting our visitors into people who stay people who engage with the site and people who find their answer on our page if we don’t do that we’re going to lose in rankings as well and so part of on-page SEO a huge part of on page SEO is satisfying the searcher that’s that is basically a conversion rate optimization activity I just a different form of it
Male Voice: Go Ahead Larry.
Larry: Oh, I was just expressing my support of what Rand just said
Male Voice: Ok, so let’s go quickly to [Unintelligible]. I think we have time for two more question. So let’s go quickly to [Unintelligible]. I think so many people have asked this all the time. The ones who are trying to start with CRO is, how long they should they run a test for?
Peep Laja: How long should we run a test for? The three factors that you should take into consideration before deciding when to stop the test. First of all you need to calculate the sample size ahead of time before you launch your test so sample size is how many people need to be part of the experiment before we can press the results and you can just google sample size calculator optimized he has one video has one there are many that independent ones and so you put in your existing baseline conversion rate for that specific page that you’re running the tests are so for the you know if it’s a your checkout page conversion rate for that page might be like ninety percent so I put in that the minimum uplift you want to be able to detect you know maybe it’s five percent whatever and they will tell you how many people you need per variation so that’s one still you should not stop the test before you have at least that many people in a variation. 2: is you need time representation test duration so you know if you’re Amazon or booking.com you can get a hundred thousand people from variation in like five minutes so is that is your tests done no no because that’s now you’re taking it convenience sample lot of representative sample of people so you need at minimum seven days’ worth of traffic because every your people on Monday behave different from people that come to your site on the Friday and maybe that one week is an outlier – so it’s better to test for at least two weeks and maybe even for four weeks so two to four weeks duration and once those two criteria hasn’t been met only then you look at your statistical significance or you know if you use patient statistics the probability of feeling better today.
Rand Fishkin: [Unintelligible] would you ever cut off a test early because the results just look too damn good
Peep Laja: Too damn good no not really but if the results look terrible then yes I don’t hesitate to kill it early but if it’s a good too good to be true oh my god then that’s like our I mean I’ve seen it too many times when a huge winner at after one week is no difference after four weeks so.
Rand Fishkin: Really?
Peep Laja: yeah
Oli Gardner: I mean every single A/B test pretty much will do this at the beginning it’ll spike they’ll be a big difference because the sample size is so small there’s a high variance and they all flat line and then they may seem to change but until you’ve gone through that experience and trusted it several times when you see that kind of behavior. you don’t want to believe it’s true
Peep Laja: be skeptical of too good to be true results in I guess probably are.
Larry Kim: It happens all the time, like 99 percent of the time. The winner becomes the loser or whatever.
Rand Fishkin: well I have a landing page right now that’s converting about one out of every Twelve thousand visits.
Rand Fishkin: I’m serious. I’m totally serious and I’m kind of hoping I have a new landing page has been in the queue to be built at Moz for the last four weeks I think it’ll be another two or four weeks before it gets launched but then I’m wondering you know if it starts winning within a day or two can I just be like okay see there we go that was for the missing
Oli Gardner: two out of twelve thousand – I’m good
Larry: it would be hard to go down from there Ron.
Rand Fishkin: It really what it is it is the shittiest landing page I have ever seen.
Oli Gardner: Donkey, donkey
Rand Fishkin: donkey by definition you know it’s a much lower former life it’s like a tardigrade
Oli Gardner: Donkey shit.
Male Voice: ok we’re going to have with this final question so I going to pitch it to you Rand but anyone pitch in because it’s interesting question I’m not sure this person had at the clearest understanding of CRO, or maybe they do after this. What line do you draw between CRO and Greyhat marketing? And is there a line?
Rand Fishkin: yeah I i certainly believe that there’s I think that the CRO can be viewed in two ways. It can be viewed as a long-term exercise in building a high value high quality customers who love your brand and want to amplify and share it and are proud to be customers or it can be used to get a lot of money quickly from people who are then very upset with you in the future and there are numerous ways that this can be done that I i would urge folks who are interested in this topic to check out a great article recently on dark patterns. Let me see if I can, dark patterns nui and yeah so this is let’s see there’s a good one from the Verge from a couple years ago called “Dark Patterns: Inside the Interface that’s Designed to Trick You”. Yeah there’s another one more recently. Oh, from the New York Times called “Websites won’t take no for an answer”, and they talk about these exploitative techniques as well I think that you know you have to decide for yourself what you’re trying to build if you are a you know an information products affiliate marketer and all you’re trying to do is maximize your one day revenue you’re probably going to go down that dark pattern path and not care about it but if you are a brand trying to build long term customer happiness and success you should be very cautious about the degree to which you increase conversion rate on a particular page or particular effort but cost yourself brand trust and happiness I you can see this in the sales world as well right I know plenty of SAS companies who salespeople are effective but so aggressive that despite their relatively high conversion rates they’ve turned eight out of ten of the people they talk to into permanent attractors for their company.
Oli Gardner: yeah travel companies will sometimes terrible not that there’s lots of dark matter what’s going on there where people are auditing going to fall to insurance for their you know and things like that which is terrible like most people don’t need travel insurance I have for else but yeah yeah.
Male Voice: ok cool so you have hit the 60 minute mark we missed out I think full response to the latest I think we missed three questions. I’m not a good time keeper even in my own personal life.
Rand Fishkin: We forgive you because you know with your Irish accent, you could say and do anything. Read us the phone book.
Male Voice: It’s the charm of the Irish.
Oli Gardner: Don’t start with the L, because you hit liar pretty quickly.
You can go over and read last year’s thread over on inbound.com. you can go to get involved in the discussion hopefully have some great speakers will kind of go there and I’m answer some questions as a reminder we are going to email the record and I to everyone do you like you have all of this information I want to just quickly say thank you very much to Pam, Larry, Oli, Peep, and Rand for joining us across the globe and this has been a great 60 minutes of a ton of interesting information I’ve really learned a lot I hope you guys learned a lot thank you very much for joining thank you everyone for joining and yeah we’re going to stop broadcasting now and I’ve done this for us hopefully we’ll go offline